conservation-v0


conservation-v0 considers a single species conservation problem in which the agent can increase the population at a cost.

Observation Space The agent observes the population at that time step.

Model Dynamics Dynamics follow a Ricker model.

Action Space The agent determines how much to increment the population.

Reward Function The agent is rewarded proportionally to the population level and penalized according to the magnitude of the action.

Conservation Gym
Conservation Gym
### conservation-v3 {data-commentary-width=400}
{=html} <div id="htmlwidget-027c2f82925dcfda976a" style="width:100%;height:auto;" class="datatables html-widget"></div> <script type="application/json" data-for="htmlwidget-027c2f82925dcfda976a">{"x":{"style":"bootstrap4","filter":"none","vertical":false,"fillContainer":false,"data":[[],[],[],[],[],[]],"container":"<table class=\"table table-striped table-hover row-border order-column display\">\n <thead>\n <tr>\n <th> <\/th>\n <th>agent<\/th>\n <th>team<\/th>\n <th>mean<\/th>\n <th>std<\/th>\n <th>ref<\/th>\n <\/tr>\n <\/thead>\n<\/table>","options":{"order":[3,"desc"],"columnDefs":[{"className":"dt-right","targets":[3,4]},{"orderable":false,"targets":0}],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
***
conservation-v3 considers a single species conservation problem that has a non-stationary model.
Observation Space The agent observes the population at that time step.
Model Dynamics Dynamics follow a May model in which a parameter, \(a\), continually changes, making it more likely for the population to collapse.
Action Space The agent determines how much to increment the population.
Reward Function The agent is rewarded proportionally to the population level and penalized according to the magnitude of the action.

Conservation Gym

conservation-v5


conservation-v5 considers a single species conservation problem that has a non-stationary model.

Observation Space The agent observes the population at that time step.

Model Dynamics Dynamics follow a May model in which a parameter, \(a\), continually changes, making it more likely for the population to collapse.

Action Space The agent determines how much to decrease the parameter \(a\) and move the system away from the tipping point.

Reward Function The agent is rewarded proportionally to the population level and penalized according to the magnitude of the action.

Conservation Gym
Conservation Gym
### conservation-v7 {data-commentary-width=400}
{=html} <div id="htmlwidget-a845d7da0e7547958bad" style="width:100%;height:auto;" class="datatables html-widget"></div> <script type="application/json" data-for="htmlwidget-a845d7da0e7547958bad">{"x":{"style":"bootstrap4","filter":"none","vertical":false,"fillContainer":false,"data":[[],[],[],[],[],[]],"container":"<table class=\"table table-striped table-hover row-border order-column display\">\n <thead>\n <tr>\n <th> <\/th>\n <th>agent<\/th>\n <th>team<\/th>\n <th>mean<\/th>\n <th>std<\/th>\n <th>ref<\/th>\n <\/tr>\n <\/thead>\n<\/table>","options":{"order":[3,"desc"],"columnDefs":[{"className":"dt-right","targets":[3,4]},{"orderable":false,"targets":0}],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
***
conservation-v7 considers a conservation problem that has a non-stationary model. It is similar to conservation-v5 but works on an ensemble of populations. By default, the replicate number is 1, so the default is equivalent to conservation-v5.
Observation Space The agent observes the populations at that time step.
Model Dynamics Dynamics follow a May model in which a parameter, \(a\), continually changes, making it more likely for the populations to collapse.
Action Space The agent determines how much to decrease the parameter \(a\) and move the systems away from their tipping points.
Reward Function The agent is rewarded proportionally to the mean population level and penalized according to the magnitude of the action.

Conservation Gym